182 THE PRINCIPLES OF SCIENCE 



SO explicable, we are then at once predisposed to see the simplest law 

 as that most probably genuine. However this may be, the same strain 

 of thought active in determining the laws we seek, and choose, seems 

 also acti\'e when laws fail unexpectedly. Always we seek to under- 

 stand such failure in terms of one or a very few "complicating factors." 



Of our theories, as of our laws, we hold simplicity the mark of truth. 

 But here the situation becomes very difficult. Using this touchstone 

 for the assay of scientific theories, all too often we obtain highly equiv- 

 ocal results. Retrospecti\^ely we see that simplicity cannot supply a 

 decisi\'e criterion of judgment, for a perfectly obvious reason: there 

 is complexity to the whole idea of simplicity. Simplicities differ in kind 

 as well as in degree. Perhaps we may distinguish the "simplicity" of 

 great explanatory appeal ( e.g., the strong sense of analogy aroused by 

 a convincing model ) from the "simplicity" of great correlative power 

 (e.g., "the glory of geometry"). Probably, however, the two categories 

 are seldom wholly distinct: e.g., is the "glory" ever without explana- 

 tory appeal? And, certainly, in judging simplicity we always face the 

 problem of somehow weighing together (if only against each other) 

 two qualitatively different species of simplicity— within which lurk 

 still further orders of complexity I examine hereafter. 



In time we learn to attach greater weight to one type of simplicity 

 than to another. But then, all the more clearly, as sigillum veri, sim- 

 plicity becomes a non-absolute criterion of judgment— an unknown 

 function of a whole shifting climate of opinion, both scientific and 

 cosmologic. Nevertheless, however difficult and insecure may be our 

 judgment, and however insufficient in the short term may be this 

 criterion, no scientist ever doubts that simplicity is a criterion for the 

 judgment of theories. Given the principle of intelligibility, that could 

 not be doubted. And, however our ideas of simplicity change in time, 

 we have consistently set a high value on that particular species of 

 simplicity we call "sameness." If for no other reason, we would prize 

 this just because we are so certain that, awaiting our discovery, there 

 is sameness in nature. Acceptance of the principle of intelligibility 

 renders inadmissible any doubt on this point. Consider: in humans 

 knowledge presupposes learning, and learning necessarily presup- 

 poses in nature some degree of sameness in time, place, and species. 

 Human knowledge thus presupposes precisely that uniformity Mill 

 regarded as the ultimate major premise of all inductions. But now the 

 principle of intelligibility asserts human capacity for knowledge of 



